Knowledge Lifecycle Breakdowns and Design Requirements for Retrieval-Grounded Agricultural Advisory Information Systems: A Systematic Review

In knowledge-intensive fields, repeated decision failures are often not due to a lack of documented or available knowledge, but rather to problems with how that knowledge is organized, retrieved, shared, and used in advisory information systems. This study examines postharvest agricultural advisory systems as an instance of this broader IS phenomenon, applying Alavi and Leidner’s Knowledge Management Lifecycle (KML) framework as an explicit theoretical lens to diagnose where and why these systems fail to support knowledge-intensive decision-making. A systematic review, guided by PRISMA principles, was conducted across 38 studies spanning knowledge management, information systems, agricultural extension, and digital advisory platforms. Three sub-questions, derived directly from KML stages, structured the synthesis: how postharvest knowledge is stored and organized and what limits its retrievability; what dissemination mechanisms are used and how they constrain contextual knowledge delivery; and to what extent existing systems support knowledge use at critical decision points. The findings reveal four mutually reinforcing systemic breakdowns across the KML: fragmented, static knowledge repositories; weak semantic retrieval mechanisms; one-way, non-interactive dissemination channels; and limited support for context-sensitive decision-making. These failures collectively explain why persistent postharvest losses in low-resource settings reflect IS deficiencies rather than knowledge unavailability. The study contributes a lifecycle-level IS diagnosis of advisory system failure and emphasizes the necessity of retrieval-grounded architectures integrating structured knowledge repositories, semantic retrieval, and multilingual natural language interfaces as first-class IS design requirements.